NONE: ONLINE-ADS>> DEBATE: 4/21/98 - measuring impressions, digest # 07
ONLINE-ADS>> DEBATE: 4/21/98 - measuring impressions, digest # 07
richard_at_tenagra.com
Tue, 21 Apr 1998 12:30:58 -0500 (CDT)
How should our industry measure an impression?
4/21/98, digest # 07
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This is special "debate" mailing from the Online Advertising
Discussion List. Anyone on the list is free to send in a response
to be considered for inclusion.
You will receive this mailing in addition to your normal
Online Ads posts/digests/reports.
web site supporting list: http://www.o-a.com/
archive: http://www.o-a.com/archive/
moderator: Richard Hoy
posting address: online-ads_at_o-a.com
subscribe/unsubscribe instructions are at the end of this message
========================================================================
------------------------------------------------------------------------
This week's Online Advertising Discussion List sponsor:
24/7 Media
Great sites. Flexible buying. Creative sponsorships.
Check out the fastest growing online network.
24/7 Media - Always Serving You http://www.247media.com
------------------------------------------------------------------------
========================================================================
Contents
Panelists' Responses:
1.) Tom Shields - NetGravity
2.) Dave Zinman - AdKnowledge
3.) Kathleen Bagley - Accipitor
Online Ads List Responses:
4.) Don Westrich - Is An Impression A Useful Measure A All?
------------------------------------------------------------------------
------------------------------------------------------------------------
Date: Mon, 20 Apr 1998 12:15:08 -0700
From: "Tom Shields" <tshields_at_netgravity.com>
In this post I am presenting a proposed standard for
counting advertising impressions and clicks. This standard
has been proposed to the IAB Media Measurement Task Force,
and is currently in discussion by the MMTF Technical
Subcommittee. I am presenting it here because it contains
some arguments that are germane to this discussion, and
because it outlines, in my opinion, the best way to count
advertising on the internet today.
Some very important points to understand, as you read this
proposal:
It is NOT final! Do not use this as a basis for
implementation yet! We have done some preliminary testing,
but I fully expect that we will need several sample
implementations from different vendors, including at least
one that is in the public domain, before we are ready to
declare it ready for standardization. Anyone who has
relevant data or is interested in helping with prototype
implementation should contact me or anyone on the MMTF
Technical Subcommittee.
It is NOT complete! This proposal only addresses image
banner ads. I fully expect that we will require additional
standards for other ad media, including interstitials, Java
ads, Javascript and HTML forms, embeddable media like
Enliven or Shockwave, etc. However, PLEASE do not let that
distract us from getting to a standard that can address the
majority of the advertising on the Internet today. If we
have to argue about every different kind of media, we'll
never get a standard done.
It is NOT perfect! There are browser bugs which make this
proposal less than perfect - these are outlined in the
proposal, and there may be more that I'm unaware of. It
does not count perfectly - users hitting "stop" and network
errors will still result in discrepancies. It will probably
result in increased traffic and increased implementation
costs for many sites. However, it is my belief that two
compliant implementations of this methodology will differ by
much less than the 50-600% differentials we're seeing today.
I believe this is a big step forward, and if we can
implement it we will convince the advertisers that we're
serious about measurement and standards.
Now that you've read the caveats, here's the proposal:
http://www.netgravity.com/standards/WD-countmethod.html
Feel free to send feedback to the Online Ads list, or to me directly.
-ts- tshields_at_netgravity.com Voice:(650) 655-4774
Tom Shields, VP, CTO http://www.netgravity.com/ Fax:(650) 655-4776
NetGravity, Inc 1700 S Amphlett Blvd, Ste 350, San Mateo, CA 94402-2715
-----------------------------POST NUMBER 2
Date: Tue, 21 Apr 1998 00:41:22 -0700
From: David Zinman <dzinman_at_adknowledge.com>
This post contains some questions that ad agencies and marketers should
ask about measurement when making purchase decisions, and explanations
as to why the questions are important. All can easily be added to an
RFP.
(1) What method do you use to calculate impressions
-- an ad insertion (when the html page is requested)
or an adview (when the ad graphic is requested)?
As I discussed in previous posts, adviews are much closer to actual
impressions, and usually a more conservative (lower) measure than ad
insertions.
If the site counts adviews, make sure to ask a follow up:
(1a) Do you use a cache confuser/counter
tag to count cached adviews?
If the site does not, then it will not be counting a signficant number
of adviews, causing click rates to be artificially augmented.
(2) When do you record the impression -- when the
page or ad request is received by the server, or
after the server fulfills the page or ad request?
Clearly, recording an impression after the server fulfills a request is
better than to do so before.
(3) Do you filter out robots?
I will spend a little more time on this question, since it has come up
quite a bit in this thread. Click counts, and even impression counts,
can be artificially increased by robots that request pages. There are
two ways to filter robots: using a robots.txt file or purging log files
of known robots. The first method is the most reliable, but both can be
used.
In the first method, the server contains a robots.txt file, which tells
well-behaved robots to avoid downloading files and generating false
impressions and clicks. The good news is that determining whether a site
uses a robots.txt file is relatively easy. You can see the alert to all
robots by just requesting that file from the server. For example, try an
AdKnowledge ad server or a Doubleclick ad server, located at:
http://ads01.focalink.com/robots.txt
or
http://ad.doubleclick.net/robots.txt
If no robots.txt file is used, then the server will usually return "file
not found". You can see an example of an ad server not using the
robots.txt file at MatchLogic:
http://ad.preferences.com/robots.txt
The second way to filter robot-generated impressions and clicks is to
purge all known robot domains from log files before generating reports
off them. Neither method is 100% effective, since ill-behaving robots
can ignore a robots.txt file, and there are many unidentified robots
that can not be caught by a program to purge log files.
(4) Will the site's ad management system work with
a third party system like AdKnowledge or others?
If the answer is "Yes", than you can use one of these third parties to
deliver all the ads for the campaign, eliminating the inconsistencies in
counting methodologies across sites. Since the third party system
delivers all the ads, it uses the same counting methodology to count all
the impressions in the campaign.
Importantly, if you are using a third party ad serving system to deliver
all your web ads, be sure and ask:
(4a) Does the third party system use a cache
confuser/counter tag (see question #1 above),
and is that tag is difficult for sites to support?
This is a CRITICAL question, for if some sites will not support it, your
counts will be inconsistent from site to site. This defeats the purpose
of using a third party system in the first place.
I hope these questions are helpful. Getting the buyers to ask smart
questions is the quickest way to get the sellers and technology vendors
to take notice.
-----------------------------POST NUMBER 3
Date: Tue, 21 Apr 1998 10:16:41 -0400
From: kbagley_at_accipiter.com (Kathleen Bagley)
Whatever comes out of this debate, Accipiter will continue
to stand behind the standards proposed by the IAB/CASIE when
it comes to this topic. That said, I still have two more
cents to throw down.
I believe that all parties have expressed pretty reasonable
opinions of 'how the industry should measure an ad
impression.' However, there is a big difference between a
reasonable method/standard to measure an ad impression and a
feasible method/standard to measure an ad impression.
Wouldn't it be lovely if technology was in a state such that
all ad impressions could be measured at the client/browser
level? I think all vendors, advertisers, and agencies would
agree that if a browser could report back to the counting
mechanism, 'Hey, I got the ad and I showed the whole thing
one time,' ah, what a wonderful world that would be.
Unfortunately, this is not feasible considering the state of
technology today. Besides, even if we could get commitment
from Netscape, Microsoft, AOL, Prodigy, and the numerous
other browser types to buy into and enforce this standard,
how can the advertiser be sure the end user 'saw' the ad?
Or, to counter that, what if an entire department of 6 is
crowded around that PC?
Surely traditional marketers have faced this dilemma in the
past? How do they handle ad measurement? I'm glad you
asked. They make assumptions and trust that, in the long
haul, it all evens out and that everyone is basing ad-buys
on an even playing field.
Similar to what I proposed in my first post regarding
measuring ads - the TV, print, and radio industries make an
assumption based on the lowest common denominator. The ad
was printed. The TV or radio station served a commercial.
I know, I know, there are million of counter arguments; the
TV/radio volume could have been turned down, the mailman
didn't deliver the magazine, the Seinfeld viewer went to the
bathroom at the first commercial break, etc. Bottom line,
all these exceptions don't really matter in the long haul.
Advertisers are dealing with these exceptions with all shows
and with all magazines. What really does matter is that
publishers, TV, and radio stations are giving advertisers an
even playing ground across all their ad buys. That's what
ultimately matters, because advertisers are the ones who are
ultimately paying. Advertisers have learned to trust that
occasionally a TV screws up and shows their ad twice.
Occasionally someone goes to the bathroom at a commercial
break. Regardless, they have an advertising metric that is
consistent across all TV shows, all publications, and all
radio stations, which gives them a level of comfort and
third-party validation of where they are spending their
money.
We can talk this until we are blue in the face - but let's
not. I honestly feel that we should not try and re-invent
the wheel here---at least not until technology matures from
its current state. Let's give the advertisers a reasonable
lowest common denominator---let the measurement of an ad
impression be counted when the ad delivery responder
successfully responds to an ad request. I printed the
publication. I showed the commercial. I served the ad. To
get back to my first point, this counting standard satisfies
the following conditions:
1. Easy to audit by 3rd Parties
2. Accurate and Reasonable to implement
3. Provide comparable data across sites that is acceptable
to advertisers and agencies
Why don't we all suggest a standard that can easily be
audited, implemented, while offering an even playing field?
I know this is not the ideal way to measure an ad
impression, but it is reasonable and feasible considering
where we are today. Now, let's get our heads back in the
game. Let's try and talk about some issues that are going
to grow the Internet pie - like getting more advertisers
online! We can get behind the IAB/CASIE with a proposed
standard and reconvene when the technology and the industry
mature.
---------------------------------------------------------
Kathleen Bagley, Public Relations and Strategic Relationships
Accipiter (http://www.accipiter.com)
Kathleen_at_accipiter.com
tel: 919-872-7755 ext. 3241
fax: 919-872-5060
Target. Deliver. Engage.
-----------------------------POST NUMBER 4
Date: Mon, 20 Apr 1998 16:46:27 -0400
From: "Don Westrich" <don_at_thethinkingmedia.com>
The Broader Question: Is An Impression A Useful Measure At
All?:
In my rant yesterday, I made the case for client-side
reporting of impressions, and why I think it's the best
technique available. However, I think we need to go back to
first principles:
The only reason advertisers care about impressions in the
first place is that it has been the best means available to
judge who sees their message, which is in turn an imperfect
means of influencing customer behavior.
"People don't buy drills, they buy holes." And buying
impressions has been a "drill" for advertisers, not the
hole.
The hierarchy of ad delivery, in order of increasing
importance to advertisers, runs like this:
1. Ads served (what most systems count) > 2. Ads
successfully transferred > 3. ads displayed (what we define
as an impression) > 4. ads viewed (per Chris Redlitz) > 5.
ads that get a click through > 6. Ads that get a more
detailed response (i.e., the surfer telling you something
substantive about themselves, types in a keyword, supplies
an E-mail address, etc.) > 6. ads that result in a sale.
Banner-based reporting handles #1 - 4 with greater accuracy
than was previously available, and represents an improvement
from the advertiser's viewpoint.
But #5 & 6 begin to transcend advertising and become direct
marketing. I submit that 5 & 6 are the grail for
advertisers.
Let's face it, impressions will remain a critical measure,
and any method that will make them more accurate and create
a standard definition is productive. But the more
functionality and interaction you can put in the banner, the
closer to being able to fulfill the advertiserís goal you
are.
Metrics like cost of customer acquisition blow the utility
of impressions away. As banners that are a true venue for
customer contact and capable of sales become more common,
impressions will become less and less relevant.
Don Westrich
Director of Business Development
Thinking Media Corp.
www.thethinkingmedia.com
Phone: 212 352 0606
Fax: 212 352 0904
34 West 17th Street
6th Floor
New York, NY 10011
-----------------------------POST NUMBER 5
-----------------------------POST NUMBER 6
-----------------------------POST NUMBER 7
========================================================================
------------------------------------------------------------------------
This week's Online Advertising Discussion List sponsor:
24/7 Media
Great sites. Flexible buying. Creative sponsorships.
Check out the fastest growing online network.
24/7 Media - Always Serving You http://www.247media.com
------------------------------------------------------------------------
========================================================================
SUBSCRIBE INSTRUCTIONS
To subscribe to the Online Advertising Discussion List,
Fill out the form at: http://www.o-a.com/
UNSUBSCRIBE INSTRUCTIONS
You may leave the Online Advertising Discussion List at
any time by sending an email message with the word UNSUBSCRIBE
(just the word) in the body or content of your message to:
online-ads-digest-request_at_o-a.com
if you get the digest version, and
online-ads-request_at_o-a.com
if you get the non-digest version.
Online Advertising Discussion List
http://www.o-a.com/
a publication of The Tenagra Corporation
http://www.tenagra.com/
